© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 92087
    Prev
    Next
    INDEX
    Explanations

    Thoroughness or validity

    np_max-act · gemini-2.0-flash

    The neuron detects intensifying descriptive words—particularly adjectives and adverbs that stress severity or thoroughness (e.g. “pressing,” “severe,” “hefty,” “thorough,” “robust”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     which
    -0.08
    мотря
    -0.07
    .documents
    -0.07
     Easily
    -0.07
    charAt
    -0.07
     expiry
    -0.07
     if
    -0.07
    .Throw
    -0.07
    Never
    -0.07
    adds
    -0.07
    POSITIVE LOGITS
    ै।↵↵
    0.06
     路
    0.06
    ");↵↵
    0.05
    ***/↵↵
    0.05
    /pkg
    0.05
     browse
    0.05
    	thread
    0.05
     treatments
    0.05
    .alibaba
    0.05
     SECRET
    0.05
    Activations Density 0.744%

    No Known Activations